On Mon, 18 Jan 2010, Eric Kow wrote:
1. One proposal that came up recently was to hide the "filler" patch by
defaulting it to 'no'; see
<http://irclog.perlgeek.de/darcs/2010-01-17#i_1907325>. I was
initially very enthusiastic about this, but after sleeping on it,
I realised that this would break hunk splitting (the filler patch
is one of the split results)
I'm not sure that this would break hunk splitting.
If you are trying to split a hunk, it's because there's a bit you want and
a bit you don't want. Because they touch each other, the newly split hunks
will have a dependency between them, and the order of that dependency is
determined by the order in which they are created, i.e. you have to keep
the bit you want in the text you edit. I think this was the confusing
thing that Mark described where he wasn't allowed to select the bit he
actually wanted.
Now, you could imagine record having some special knowledge of touching
hunks and allowing them to commute just for the purposes of selection. But
that's a road I'm very reluctant to go down as it would be quite messy and
special-cased.
So given this, having the filler patch default to no still makes sense to
me, because even in the hunk splitting case, it's by definition the bit
you don't want.
BTW I do have an implementation of this ready and I'll send it in shortly
just so it's available to try out.
Ganesh
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users
